According to sources cited by The Information, OpenAI’s upcoming AI model, referred to internally as Orion, may not demonstrate the same significant advancements as its predecessors. Although it outperforms current models, the incremental progress from GPT-3 to GPT-4 is not as noticeable with Orion, especially in specialized areas such as coding. This indicates a potential slowdown in the rate of AI advancements.

In response to these challenges, OpenAI has initiated a new strategy by forming a dedicated foundations team. This team is exploring innovative techniques to enhance model training, focusing on the use of synthetic data generated by AI and refining models post-training to push the boundaries of current AI capabilities.

Despite these developments, OpenAI has not confirmed the release of a model named Orion this year. Previously, the company addressed speculation about its flagship models by clarifying its release plans, underscoring a cautious approach to the ongoing evolution and optimization in the field of artificial intelligence.